Search

Recent cyber-physical attacks, such as Stuxnet, Triton etc., have invoked an ominous realization about the vulnerability of critical infrastructure, including water, power and gas distribution systems. Traditional IT security-biased protection methods that focus on improving cyber hygiene are largely impotent in the face of targeted attacks by advanced cyber-adversaries. Thus, there is an urgent need to analyze the safety and security of critical infrastructure in a holistic fashion, leveraging the physics of the cyber-physical system. System-Theoretic Accident Model & Processes (STAMP) offers a powerful framework to analyze complex systems; hitherto, STAMP has been used extensively to perform safety analyses but an integrated safety and cybersecurity analysis of industrial control systems (ICS) has not been published. This paper uses the electrical generation and distribution system of an archetypal industrial facility to demonstrate the application of a STAMP-based method called Cybersafety to identify and mitigate cyber- vulnerabilities in ICS. The key contribution of this work is to differentiate the additional steps required to perform a holistic cybersecurity analysis for an ICS of significant size and complexity and to present the analysis in a structured format that can be emulated for larger systems with many interdependent subsystems.

With the growth of mobile data traffic in wireless networks, caches are used to bring data closer to mobile users and to minimise the traffic load on macro base station (MBS). Storing data in caches on user terminals (UTs) and small base stations (SBSs) faces challenges with respect to the decision of cache contents. Here, a multi‐objective cache content strategy that aims to maximise the cache hitrate of SBSs in mobile edge networks (MENs) is proposed. The multi‐objective cache placement optimisation is formulated as a classification problem. Unlike previous work, mobility input attributes such as user locations, contact duration, communication ranges, contact probability between UTs and SBSs, etc. as well as content popularity and the correlation between these input attributes separating the decision space into two regions of cache and not cache are used.Stochastic gradient descent algorithm is used for the training of three supervised machine learning techniques: artificial neural network ANN, support vector machine (SVM), and logistic regression LR to define the hyperplane that separates the cache content decision space. Simulation results show that compared with the weighted‐sum approach, the SBSs cache hit rates increase on the average by 18.58%, 18.52%, and 18.2%, and the total energy consumption values decrease on the average by 33.49%, 53.19%, and 49.9% for ANN, SVM, and LR, respectively.

In this paper, we propose a solution to the problem of scheduling of a smart home appliance operation in a given time range. In addition to power-consuming appliances, we adopt a photovoltaic (PV) panel as a power-producing appliance that acts as a micro-grid. An appliance operation is modeled in terms of uninterruptible sequence phases, given in a load demand profile with a goal of minimizing electricity cost fulfilling duration, energy requirement, and user preference constraints. An optimization algorithm, which can provide a schedule for smart home appliance usage, is proposed based on the mixed-integer programming technique. Simulation results demonstrate the utility of our proposed solution for appliance scheduling. We further show that adding a PV system in the home results in the reduction of electricity bills and the export of energy to the national grid in times when solar energy production is more than the demand of the home.

Over the past few years, automatic recognition of human interactions has drawn significant attention from researchers working in the field of Artificial Intelligence (AI). And feature extraction is one of the most critical tasks in developing efficient Human Interaction Recognition (HIR) systems. Moreover, recent researches in computer vision suggest that robust features lead to higher recognition accuracies. Hence, an improved HIR system has been proposed in this paper that combines 2D and 3D features extracted using machine learning and deep learning techniques. These discriminative features result in accurate classification and help avoid misclassification of similar interactions. Ten keyframes have been extracted from each video to reduce computational complexity. Next, these frames have been preprocessed using image normalization and noise removal techniques. The Region Of Interest (ROI), which contains the two humans involved in the interaction, has been extracted using motion detection. Then, the human silhouettes have been segmented using the GrabCut algorithm. Next, the extracted silhouettes have been converted into 3D meshes and their heat kernel signatures (HKS) have been obtained to extract key body points. A Convolutional Neural Network (CNN) has been used to extract full-body features from 2D full-body silhouettes. Then, topological and geometric features have been extracted from the key body points. Finally, the combined feature vector has been fed into Long Short-Term Memory (LSTM) and each interaction has been recognized using a Softmax classifier. The proposed system has been validated via extensive experimentation on three challenging RGB+D datasets. The recognition accuracies of 91.63%, 90.54%, and 90.13% have been achieved with the SBU Kinect Interaction, NTU RGB+D, and ISR-UoL 3D social activity datasets respectively. The results of extensive experiments performed on the proposed system suggest that it can be used effectively for various applications, such as security, surveillance, health monitoring, and assisted living.

In this paper, we evaluate and analyze the performance of long short-term memory networks (LSTMs) for individual electric vehicle (EV) charging behavior prediction over the next day. The charging behavior consists of the charging duration level within a certain upper and lower range, the time slots in which charging will take place, the number of times charging will take place in each time slot, and whetherthenextdaywillbeachargingdayornot.Unlikeexistingwork,weevaluatethebehaviorprediction performance for increasing resolutions of charging duration levels and charging time slots, using varying lengths of training data. The performance of the proposed approach is validated using real EV charging data, and comparison with other machine learning methods shows its generally superior prediction accuracy for all resolutions. We show that the best performance is achieved when around 8-10 months of data are used as training data. It is also shown that although the performance of the LSTMs degrades with increasing resolution, the performance for charging time slot prediction is affected less compared to that for charging duration prediction. We further propose, analyze and evaluate a new technique that improves the charging duration prediction performance.

In this article, we present a data obfuscation technique for smart meter data based on additive correlated noise. This noise is used to mask the data transmitted by different users to a third party, resulting in protection against eavesdroppers, but at the same time enabling the accurate recovery of statistics of the original data for use by the energy supplier. We analyze the proposed technique by studying its obfuscation performance and accuracy of statistics recovered from the masked data as a function of noise correlation with the users’ data. Finally, we identify deep learning techniques such as generative adversarial networks for data obfuscation using correlated noise, and show preliminary results to demonstrate their performance in this scenario.

A nano-grid is an independent hybrid sustainable framework that utilizes non-renewable and renewable power resources for supplying continuous electrical energy to the load. Considering this scenario, in this research work, photovoltaic (PV) array, wind turbine, and fuel cell are taken as the three generation resources that have been used in the nano-grid. The active and reactive power of the all three generation resources is controlled using various controllers, i.e. integral, proportional-integral, proportional derivative, proportional integral derivative, fractional-order proportional-integral, fractional order proportional integral derivative (FOPID) and sliding mode controller (SMC). An advanced optimization technique based on a genetic algorithm (GA) and particle swarm optimization (PSO) algorithm has been utilized to optimize all of these controllers. The integral square error is taken as the objective function for both optimization algorithms. Finally, a graphical and tabular comparative analysis of all optimized controllers along with their control parameters and performance indexes is evaluated to find the best optimal solution. The performance of SMC has surpassed the performance of all other optimized controllers for power stability. In less than 0.267 seconds, the actual power yielded by using SMC is within 1% of the desired power. PSO algorithm has performed better than GA algorithm with all controllers. The worst performance is by FOPID controller with a steady state error of 6071.3W using GA algorithm and have a high magnitude of overshoot and undershoot. Moreover, a smart switching algorithm has been introduced for switching between the generation resources in accordance with the load demand and cost of the system in order to operate the nano-grid more economically. Finally, a case study has been performed in which the smart switching algorithm is utilized to switch to the best available generation resource in case of any fault at the generation side to provide uninterrupted power to the attached loads.

Physical control systems are increasingly controlled by reconfigurable, network-enabled devices to increase flexibility and ease commissioning and maintenance. Such capability creates vulnerabilities. Devices may be remotely reprogrammed by a malicious actor to act in unintended ways, causing physical damage to mechanical equipment, infrastructure, and life and limb. In this paper, past examples of actual damage to cyber-physical systems are shown, threats posed by software-controlled variable frequency drives (VFDs) are analyzed, and a small-scale version of an attack on ubiquitous VFD equipment is demonstrated.

Background: In recent decades, enormous efforts for different drug discovery processes have led to a number of drug molecules available today to overcome different challenges of the health care system. Unfortunately, more than half of these drugs are listed in either BCS (biopharmaceutical classification system) class II/ IV or both are eliminated from the development pipeline due to their limited clinical use. A nanotechnological approach bears much hope and lipoidal fabrication is found to be suitable for the delivery of such drugs. Nanoemulsion based gel i.e. nanoemulgel out of different nanolipoidal formulations has been found to be a suitable approach to successful drug delivery through topical routes. In past few years many herbal and synthetic active pharmaceutical ingredients (APIs) has been patented as nano sized emulsified gel for various therapeutic activities. Methods: Nanoemulgel is basically an emulsion-based topical gel formulation, where nanosized emulsion globules can be prepared with the help of high energy or low energy methods and further converted into nanoemulgel by adding a suitable gelling agent. Nanoemulgel fabrication enlists various kinds of polymeric materials, surfactants and fatty substances of natural, synthetic and semi-synthetic nature with a globule size range from 5 to 500 nm. Results: Nanoemulgel can be applicable to various acute and chronic diseases through topical routes. Conclusion: Nanoemulgel preparations of many recently approved drugs are being used successfully in different areas of health care and have re-defined the significance of topical route of delivery as compared to other routes. However, along with various improvements in the current state of the delivery system, the safety factor needs to be taken into account by toxicological studies of the materials used in such formulations.

Time-series reconstruction in differential interferometric synthetic aperture radar (DInSAR) requires solving a regression problem and can be used to monitor surface deformation. In this study, the authors present an improved time-series reconstruction for DInSAR using Huber-norm minimisation. The Huber-norm combines l(2)-norm and l(1)-norm based on a threshold retaining features of both norms. Unlike existing techniques based on either l(2)-norm or l(1)-norm minimisation, their proposed technique has better reconstruction performance in the presence of both outlier-like phase unwrapping errors and additive Gaussian noise, as a Huber density can model these errors. They first validate the proposed technique using real European Remote Sensing satellite data. To further analyse the performance, they simulate phase data according to different deformation models with phase unwrapping errors and additive Gaussian noise. They show via Monte-Carlo simulations that their method has superior performance, compared to the existing approaches. They further verify the performance of the proposed technique with respect to the threshold parameter and for a varying percentage of phase unwrapping errors.

In this paper, we propose an environment perception framework for autonomous driving using state representation learning (SRL). Unlike existing Q-learning based methods for efficient environment perception and object detection, our proposed method takes the learning loss into account under deterministic as well as stochastic policy gradient. Through a combination of variational autoencoder (VAE), deep deterministic policy gradient (DDPG), and soft actor-critic (SAC), we focus on uninterrupted and reasonably safe autonomous driving without steering off the track for a considerable driving distance. To ensure the effectiveness of the scheme over a sustained period of time, we employ a reward-penalty based system where a higher negative penalty is associated with an unfavourable action and a comparatively lower positive reward is awarded for favourable actions. The results obtained through simulations on DonKey simulator show the effectiveness of our proposed method by examining the variations in policy loss, value loss, reward function, and cumulative reward for ‘VAE+DDPG’ and ‘VAE+SAC’ over the learning process.

With the growth of mobile data traffic in wireless networks, caches are used to bring data closer to mobile users and to minimize the traffic load on macro base station (MBS). Storing data in caches on user terminals (UTs) and small base stations (SBSs) faces challenges on which data to cache and where to cache these data. The process of deciding the cache contents involves multiple objectives regarding the content popularity, contact duration between UT and SBSs, communication ranges between UT and SBSs caches, and contact probability between UT and SBSs. In this paper, we propose a new strategy on cache placement decisions for mobile edge networks based on binary classification technique. The aim is to formulate the cache placement as a classification problem that is solved using machine learning techniques in order to define an optimal decision boundary on cache or not cache decisions. Simulation results show that the performance of cache placement algorithms using classifier based learning techniques can achieve higher hit rate than other algorithms.

This work presents the grouping of dependent tasks into a cluster using the Bayesian analysis model to solve the affinity scheduling problem in heterogeneous multicore systems. The non-affinity scheduling of tasks has a negative impact as the overall execution time for the tasks increases. Furthermore, non-affinity-based scheduling also limits the potential for data reuse in the caches so it becomes necessary to bring the same data into the caches multiple times. In heterogeneous multicore systems, it is essential to address the load balancing problem as all cores are operating at varying frequencies. We propose two techniques to solve the load balancing issue, one being designated “chunk-based scheduler” (CBS) which is applied to the heterogeneous systems while the other system is “quantum-based intra-core task migration” (QBICTM) where each task is given a fair and equal chance to run on the fastest core. Results show 30–55% improvement in the average execution time of the tasks by applying our CBS or QBICTM scheduler compare to other traditional schedulers when compared using the same operating system.

Problems in the shared caches in multicore systems arise due to the non-affinity scheduling. Tasks are scheduled without considering the possible dependencies they have on each other. It has a negative effect on the overall execution time of the tasks. In this paper, we have proposed affinity based scheduling using Bayesian analysis model and creating groups or clusters of dependent tasks. Clusters are then allocated fairly and equally among the multiple cores. Load balancing is performed on the homogeneous system by feeding all the cores in a multicore architecture from a queue-like pool of tasks. We have used another technique for load balancing by defining a chunk size for each core. Results showed an improvement in an overall execution time of a process by 5.57% and of an individual task by 9.06% on average in comparison with other traditional schedulers used by the operating system for a factorial program. For a quick sort program, overall execution time of a process has been reduced by 1.13% while for an individual task by 1.5%.

The sole purpose of providing education, whether primary or advanced, is to impart knowledge that can be applied to real-life issues. This process inherently requires finding the degree of knowledge acquired by the learner, which can be achieved by conducting examinations of various types. This is then followed by a uniform assessment via examination. In an evaluation by descriptive examination, students are assessed at a higher level of Bloom’s Taxonomy as compared to an objective examination. Unfortunately, evaluation in a descriptive examination is a cumbersome process and is done manually due to the complexity and ambiguity present in natural language. This work is an effort to deal with the problem of automated computer assessment in descriptive examination and presents a step-by-step algorithm for the programmed evaluation of descriptive answers by considering both syntactic and semantic similarities. Conceptual similarity and word sense disambiguation are also considered for better results in the evaluation process. The proposed model produces convincing results and is very comparable to human evaluation.

A newfangled family of Nyquist-I pulses is proposed and named improved dual sinc pulse (IDSP). The IDSP is designed to improve performance in orthogonal frequency division multiplexing (OFDM)-based systems. The IDSP is a generalization of the dual sinc pulse (DSPP). This is because the DSP was formulated for α = 1 whereas the IDSPP is valid for 0 ≤ α ≤ 1. The behavior of the IDSP is promising in terms of its frequency and time domain responses. Theoretical and numerical outcomes indicate that the IDSP outperformed other existing pulses applied in OFDM-based systems for various key evaluation metrics.

Future mobile services and applications are bounded by user location, data, and network. These services will suffer from poor support from wireless networks due to the huge amount of mobile traffic and user mobility. The demand for contents by these services results in constraints put on latency and quality of service (QoS). Considering these problems, researchers investigated caching contents locally and proactively at the edge of the mobile edge networks (MENs). In this work, we proposed new formulation of mobility-aware latency-efficient cache placement problem for mobile edge networks (MENs) taking into account different storage capacities, users mobility, content popularity, contact probability, and latency to download the contents to user terminals (UTs). Our formulated multi-objective optimization problem aims to maximize the cache hit rate. We apply weighted-sum decision theory approach to model the decision of placing contents at the edge of the network. Simulation results are shown to gain insight of the impact of different factors on our proposed system and and evaluated the results with three other cache placement techniques.

Production of energy and its usage per capita is considered as indicators that depict the living standards of a nation. The economic growth of a country is directly dependent on the availability of energy resources. To create a balance between energy demand and available energy resources, renewable resources are being in much focused these days. The most important renewable energy resources include solar, wind, geothermal and biogas, etc. Solar energy is one of the most important resources of renewable energy, used to convert light energy directly into electrical energy. Maximum power extraction is a major problem associated with Photovoltaic (PV) power systems. Maximum power point tracking (MPPT) is required to improve the energy conversion efficiency of, Photovoltaic power system. The MPPT is a dynamic process of getting the optimal output power from solar PV Arrays under different irradiation levels and temperature change. The performance of an MPPT Controller is dependent on the fact that how rapidly it can attain the Maximum Power Point (MPP) and stability of output. Previously Different MPPT Controllers have been designed to extract maximum power from PV modules under changing weather conditions. To overcome the limitations of previous methods, an intelligent method based on fuzzy logic is proposed for maximum power point tracking. To analyze the performance of fuzzy logic based MPPT controller a PV System is designed in MATLAB/SIMULINK. The resulted Simulations show that the proposed system has a fast response and good performance characteristics. It can improve the efficiency of PV array improves effectively against change in irradiation and temperature.

In this research, a fault-tolerant control scheme is designed and implemented to estimate and increase the performance of the wind turbine by using the Linear Quadratic Regulator (LQR) controller. LQR is an optimal control regulator that tracks the reference trajectory at a higher efficiency rate as compared to traditional controllers such as a PI controller, which has been used in the literature. The standard model of a wind turbine with two actuator faults is implemented in Simulink. In order to enhance the reliability and stability of the wind turbine system, the values are considered and the controller design approach is modeled and executed in the pitch system. The system is executed in the presence of actuator faults, which are the air content in oil and hydraulic leakage. Intensive simulations have been performed, to test the feasibility of the designed control scheme, on the benchmark wind turbine model. The efficiency of the wind turbine system has been successfully increased by removing the two faults of the actuator.

Availability of Energy resources is very important for economic growth of a country. Conventional energy resources are exhaustible and will deplete in near future, so it is need of hour to find new sources of energy. To meet current energy demand renewable resources especially Solar Photovoltaic (SPV) energy is being in much focus. If a PV system generates output that is more than the load requirement. The produced excessive energy will cause reverse current flowing towards the generator causing damage. For grid tied PV systems multiple MPPT inverter can be used to avoid this reverse feed current. Multiple MPPT control algorithm works by controlling the output of system by adjusting Maximum Power Point using feedback from generator. A grid-tied PV system connected with synchronous generator is designed in MATLAB Simulink toll to investigate the performance of proposed MMPT Control Algorithm for controlling Reverse Feed Current (RFC).

In this paper, we propose an environment perception framework for autonomous driving using state representation learning (SRL). Unlike existing Q-learning based methods for efficient environment perception and object detection, our proposed method takes the learning loss into account under deterministic as well as stochastic policy gradient. Through a combination of variational autoencoder (VAE), deep deterministic policy gradient (DDPG), and soft actor-critic (SAC), we focus on uninterrupted and reasonably safe autonomous driving without steering off the track for a considerable driving distance. Our proposed technique exhibits learning in autonomous vehicles under complex interactions with the environment, without being explicitly trained on driving datasets. To ensure the effectiveness of the scheme over a sustained period of time, we employ a reward-penalty based system where a negative reward is associated with an unfavourable action and a positive reward is awarded for favourable actions. The results obtained through simulations on DonKey simulator show the effectiveness of our proposed method by examining the variations in policy loss, value loss, reward function, and cumulative reward for `VAE+DDPG&rsquo, and `VAE+SAC&rsquo, over the learning process.

Mobile sensors are playing a vital role in various applications of a normal day life. Key size in securing data is an important issue to highlight in mobile sensor data transfer between a smart device and a data storage component. Such key size may affect memory storage and processing power of a mobile device. Therefore, we proposed a secure mobile sensor data transfer protocol called secure sensor protocol (SSP). SSP is based on Elliptic Curve Cryptography (ECC), which generates small size key in contrast to conventional asymmetric algorithms like RSA and Diffie Hellman. SSP receive values from light sensor and magnetic flux meter of a smart device. SSP encrypts mobile sensor data using ECC and afterwards it stores cipher information in MySQL database to receive remote data access. We compared the performance of the ECC with other existing asymmetric cryptography algorithms in terms of secure mobile sensor data transfer based on data encryption and decryption time, key size and encoded data size. In-addition, SSP shows better results than other cryptography algorithms in terms of secure mobile sensor data transfer.

In this paper, we present a novel approach for individual electric vehicle (EV) charging prediction based on long short-term memory networks. Unlike existing methods, our proposed approach does not require any charging information from the other EV users, and can also separately predict the total charging duration on the next day within a certain range. In addition, the proposed approach can forecast the charging start time slots for the next day, and the number of times charging will take place in each time slot. This information can also be used to predict whether the next day will be a charging day or not. The performance of the proposed approach is validated using real EV charging data, and comparison with other machine learning methods shows its superior prediction accuracy.

Orthogonal frequency division multiplexing (OFDM) is sensitive to frequency offset since it causes inter-carrier interference (ICI). Frequency offset results from the receiver's crystal oscillator inaccuracy, Doppler spread, or distortion. The pulse shaping function technique has proven its potential to combat ICI. To enhance the OFDM tolerance against the frequency offset, we introduce the improved double-jump 1 filter. The novel filter is then optimized given a representative system configuration. We finally demonstrate the effectiveness of the proposed filter in regard to the other novel pulses. This work serves to increase the wireless traffic demand.

Assessment of human behavior during performance of daily routine actions at indoor areas plays a significant role in healthcare services and smart homes for elderly and disabled people. During this consideration, initially, depth images are captured using depth camera and segment human silhouettes due to color and intensity variation. Features considered spatiotemporal properties and obtained from the human body color joints and depth silhouettes information. Joint displacement and specific-motion features are obtained from human body color joints and side-frame differentiation features are processed based on depth data to improve classification performance. Lastly, recognizer engine is used to recognize different activities. Unlike conventional results that were evaluated using a single dataset, our experimental results have shown state-of-the-art accuracy of 88.9% and 66.70% over two challenging depth datasets. The proposed system should be serviceable with major contributions in different consumer application systems such as smart homes, video surveillance and health monitoring systems.

Pulse shaping is used in orthogonal frequency division multiplexing (OFDM) based systems to reduce inter-carrier interference (ICI) power and peak-to-average power ratio (PAPR), which are considered the major weaknesses in OFDM-based systems. A novel family of Nyquist-I pulses called sinc exponential pulse (SEP) is proposed, and it is characterized by two new design parameters that provide extra degrees of freedom for a certain roll-off factor, ź. SEP effectively decreases the relative magnitude of the two largest side lobes of the SEP frequency function, which minimizes the ICI power and reduces the PAPR in OFDM systems. Furthermore, the SEP possesses a broader main lobe, which provides sufficient improvement in bit-error-rate (BER). The behavior of the SEP is examined in the time and frequency domain by tuning its design parameters to obtain the sub-optimum SEP. Theoretical and simulation results show that the sub-optimum SEP performs better than other existing pulses in terms of ICI power, signal-to-interference ratio (SIR) power, BER, and PAPR in OFDM-based systems.

Recent world events and geopolitics have brought the vulnerability of critical infrastructure to cyberattacks by advanced cyber-adversaries to the forefront. While there has been continued and growing attention to vulnerabilities in our energy generation and distribution systems, there are many other types of industrial control systems, ranging from manufacturing facilities to refineries to water treatment plants, that are also at significant risk based on their unique physical characteristics. In this paper, we take a holistic view of one such complex system – the chiller plant – in an archetypal industrial setting to identify vulnerabilities emerging from interactions between technology, operator actions as well as organizational structure and provide recommendations to mitigate resulting loss scenarios.

Nowadays the vast amount of video data is generated, which needs to be processed fast for getting correct insights into it. Manpower is not enough to process all of it in a short span of time due to their limited capacity. Video analytics facilitates useful tools to extract more valuable video data. This paper lays stress on different videos analysis activities. Here we explore various techniques which are used in video analysis along with presenting a comparison with image analysis techniques. Further, the application of video analytics along with the challenges being faced while analyzing the real time videos is discussed.

Power quality has become a significant concern for the Electrical Engineers over the years. Among the numerous power quality issues, the most distinguished is the voltage sag. In this research, Pulse-Width Modulation (PWM) based switched autotransformer is used in order to recompense the depletion in voltage. This technique utilizes reduced number of switching devices per phase, lessened gate driver circuit size, the ability to preserve the desired undistorted load voltage, wholesomeness and quick response. Therefore, it is inferred that the load voltage is sustained in spite of the sag.

Global positioning system (GPS) tracks a desired location and display it on map. However, GPS may not work properly for indoor positioning, when we are concerned with an accurate location of a person inside a building. Therefore, to determine an accurate indoor positioning based on smart devices a scheme is proposed and called iLocation. iLocation measures the location by obtaining and computing values from an accelerometer and a barometer of a smart device. iLocation uses MySQL to maintain records of the computed indoor positions. Such positions are displayed on Google Maps with its images for the verification. iLocation is tested on various real time scenarios and experimental results shows the accuracy, which is better than other existing approaches.

Orthogonal frequency division multiplexing (OFDM) is highly sensitive to frequency offset since it induces inter-carrier interference (ICI). Fortunately, the pulse shaping filter method has proven its potential to reduce ICI without the cost of increasing the complexity of the OFDM modem. In order to improve the OFDM tolerance against the oscillator frequency detuning, we introduce the better than double jump 2 (BTDJ2) pulse. It is characterized by one additional design parameter apart from the roll-off factor. In addition, the improved filter is optimized following a representative system configuration. The BTDJ2 is compared with other novel pulses in terms of signal to noise ratio (SNR) and frequency offset requirements for a well known practical forward error correction (FEC) limit. Further, complexity evaluation of the BTDJ2 and other novel pulses is done in terms of the average elapsed convolution time. In general, the sub-optimum BTDJ2 outperforms other novel pulses in terms of the bit error rate (BER) and complexity.

This article presents a comprehensive survey of deep learning applications for object detection and scene perception in autonomous vehicles. Unlike existing review papers, we examine the theory underlying self-driving vehicles from deep learning perspective and current implementations, followed by their critical evaluations. Deep learning is one potential solution for object detection and scene perception problems, which can enable algorithm-driven and data-driven cars. In this article, we aim to bridge the gap between deep learning and self-driving cars through a comprehensive survey. We begin with an introduction to self-driving cars, deep learning, and computer vision followed by an overview of artificial general intelligence. Then, we classify existing powerful deep learning libraries and their role and significance in the growth of deep learning. Finally, we discuss several techniques that address the image perception issues in real-time driving, and critically evaluate recent implementations and tests conducted on self-driving cars. The findings and practices at various stages are summarized to correlate prevalent and futuristic techniques, and the applicability, scalability and feasibility of deep learning to self-driving cars for achieving safe driving without human intervention. Based on the current survey, several recommendations for further research are discussed at the end of this article.

The development of disaster preparedness and forecasting system integrating real-time sensor data and time stringent computing is becoming increasingly important for mitigating the negative effects of disasters. In this paper, a service oriented IoT architecture for an early warning system (EWS) through implementation of machine learning (ML) algorithm in a cloud server is presented. Unlike other EWS approaches, our implementation includes layer level IoT architecture with a triple modular redundancy fault tolerant scheme for sensor network to guarantee availability of reliable and fault free sensor data to the cloud for accurate predictions. A ML algorithm is implemented for three sensor parameters viz., MQ4, MQ7 and force sensing resistor on the AWS cloud. The absolute error percentage between the actual and the predicted values are found to be 6.18%, 3.03% and 3.65%, respectively, for a set of values at 20 different time intervals.

Based on the rapid increase in the demand for people counting and tracking systems for surveillance applications, there is a critical need for more accurate, efficient, and reliable systems. The main goal of this study was to develop an accurate, sustainable, and efficient system that is capable of error-free counting and tracking in public places. The major objective of this research is to develop a system that can perform well in different orientations, different densities, and different backgrounds. We propose an accurate and novel approach consisting of preprocessing, object detection, people verification, particle flow, feature extraction, self-organizing map (SOM) based clustering, people counting, and people tracking. Initially, filters are applied to preprocess images and detect objects. Next, random particles are distributed, and features are extracted. Subsequently, particle flows are clustered using a self-organizing map, and people counting and tracking are performed based on motion trajectories. Experimental results on the PETS-2009 dataset reveal an accuracy of 86.9% for people counting and 87.5% for people tracking, while experimental results on the TUD-Pedestrian dataset yield 94.2% accuracy for people counting and 94.5% for people tracking. The proposed system is a useful tool for medium-density crowds and can play a vital role in people counting and tracking applications.

This paper presents an improved technique for short-term electric load forecasting making use of boosted neural networks (BooNN). The BooNN consist of combining a set of artificial neural networks (ANNs) trained iteratively. At each iteration, the error between the estimated output from the ANN model trained in the previous iteration and the target output is minimized. The final predicted result is the weighted sum of output from all the trained models. This process reduces the magnitude of forecasting errors and their variation compared to a single ANN and bagged neural networks (BNN). It further significantly lowers computational time compared to BNN. Results with real data further confirm that BooNN lead to improved load forecasting performance with respect to other existing techniques.

The recently developed depth imaging technologies have provided new directions for human activity recognition (HAR) without attaching optical markers or any other motion sensors to human body parts. In this paper, we propose novel multi-fused features for online human activity recognition (HAR) system that recognizes human activities from continuous sequences of depth map. The proposed online HAR system segments human depth silhouettes using temporal human motion information as well as it obtains human skeleton joints using spatiotemporal human body information. Then, it extracts the spatiotemporal multi-fused features that concatenate four skeleton joint features and one body shape feature. Skeleton joint features include the torso-based distance feature (DT), the key joint-based distance feature (DK), the spatiotemporal magnitude feature (M) and the spatiotemporal directional angle feature (θ). The body shape feature called HOG-DDS represents the projections of the depth differential silhouettes (DDS) between two consecutive frames onto three orthogonal planes by the histogram of oriented gradients (HOG) format. The size of the proposed spatiotemporal multi-fused feature is reduced by a code vector in the code book which is generated by vector quantization method. Then, it trains the hidden Markov model (HMM) with the code vectors of the multi-fused features and recognizes the segmented human activity by the forward spotting scheme using the trained HMM-based human activity classifiers. The experimental results on three challenging depth video datasets such as IM-DailyDepthActivity, MSRAction3D and MSRDailyActivity3D demonstrate that the proposed online HAR method using the proposed multi-fused features outperforms the state-of-the-art HAR methods in terms of recognition accuracy.

Human activity recognition using depth information is an emerging and challenging technology in computer vision due to its considerable attention by many practical applications such as smart home/office system, personal health care and 3D video games. This paper presents a novel framework of 3D human body detection, tracking and recognition from depth video sequences using spatiotemporal features and modified HMM. To detect human silhouette, raw depth data is examined to extract human silhouette by considering spatial continuity and constraints of human motion information. While, frame differentiation is used to track human movements. Features extraction mechanism consists of spatial depth shape features and temporal joints features are used to improve classification performance. Both of these features are fused together to recognize different activities using the modified hidden Markov model (M-HMM). The proposed approach is evaluated on two challenging depth video datasets. Moreover, our system has significant abilities to handle subject"s body parts rotation and body parts missing which provide major contributions in human activity recognition.

Device-to-device (D2D) communications can help in achieving the higher data rate targets in emerging wireless networks. The use of D2D communication imposes certain challenges such as interference with the cellular and D2D users. A well-designed joint admission control, network mode selection, and power allocation technique in a cellular network with D2D capability can improve overall throughput. The proposed technique jointly maximizes the total throughput and number of admitted users in cellular networks under quality-of-service (QoS) and interference constraints. The joint admission control, mode selection, and power allocation problem (JACMSPA) falls into a class of mixed-integer nonlinear constraint optimization problems that are generally NP-hard. Due to the combinatorial nature of the problem, its optimal solution needs exhaustive search of integer variables whose complexity increases exponentially with the number of user pairs. In this paper, we invoke outer approximation approach (OAA)-based linearization technique to solve the JACMSPA. The proposed method gives guaranteed $\varepsilon$ - optimal solution with reasonable computational complexity. Simulation results verify the effectiveness of the proposed approach method.

This paper presents an energy-efficient cooperative MAC (EECO-MAC) protocol using power control in mobile ad hoc networks. Cooperative communications improve network performance by taking full advantage of the broadcast nature of wireless channels. The power control technique improves the network lifetime by adjusting the transmission power dynamically. We propose the best partnership selection algorithm, which takes energy consumption into consideration for selection of the optimal cooperative helper to join in the transmission. Through exchanging control packets, the optimal transmission power is allocated for senders to transmit data packets to receivers. In order to enhance energy saving, space---time backoff and time---space backoff algorithms are proposed. Simulation results show that EECO-MAC consumes less energy and prolongs the network lifetime compared to IEEE 802.11 DCF and CoopMAC at the cost of delay. Performance improvement offered by our proposed protocol is apparent in congested networks where nodes have low and limited energy.

Behavior monitoring and classification is a mechanism used to automatically identify or verify individual based on their human detection, tracking and behavior recognition from video sequences captured by a depth camera. In this paper, we designed a system that precisely classifies the nature of 3D body postures obtained by Kinect using an advanced recognizer. We proposed novel features that are suitable for depth data. These features are robust to noise, invariant to translation and scaling, and capable of monitoring fast human bodyparts movements. Lastly, advanced hidden Markov model is used to recognize different activities. In the extensive experiments, we have seen that our system consistently outperforms over three depth-based behavior datasets, i.e., IM-DailyDepthActivity, MSRDailyActivity3D and MSRAction3D in both posture classification and behavior recognition. Moreover, our system handles subject's body parts rotation, self-occlusion and body parts missing which significantly track complex activities and improve recognition rate. Due to easy accessible, low-cost and friendly deployment process of depth camera, the proposed system can be applied over various consumer-applications including patient-monitoring system, automatic video surveillance, smart homes/offices and 3D games.

Physical control systems are increasingly controlled by reconfigurable, network-enabled devices to increase flexibility and ease commissioning and maintenance. Such capability creates vulnerabilities. Devices may be remotely reprogrammed by a malicious actor to act in unintended ways, causing physical damage to mechanical equipment, infrastructure, and life and limb. In this paper, past examples of actual damage to cyberphysical systems are shown, threats posed by software-controlled Variable Frequency Drives (VFDs) are analyzed, and a small-scale version of an attack on ubiquitous VFD equipment is demonstrated.

Nowadays, there arise a tremendous increase in the amount of data or we can say the data coming from multiple sources gather to form huge data. There are various techniques used in soft computing for Multi-Source Data analysis and its technological advancements. This paper, describes an overview of multisource data analytics, its application, advantages, and limitations and few models have reviewed with their objective, merits and so on. Moreover, there are various issues in existing storage techniques and improvements in technology. Since the existing and traditional technologies have their own issues to manage huge data, so advance technologies have been derived to manage them and to derive useful insights. Further, few research issues and future directives are presented in this paper.

In this thesis, we present a new joint image enhancement and reconstruction method and a software processing tool for SAR Interferometry. First, we propose a sparsity-driven method for coupled image formation and autofocusing based on multi-channel data collected in interferometric synthetic aperture radar (IfSAR). Relative phase between SAR images contains valuable information. For example, it can be used to estimate the height of the scene in SAR Interferometry. However, this relative phase could be degraded when independent enhancement methods are used over SAR image pairs. Previously, Ramakrishnan, Ertin, and Moses proposed a coupled multi-channel image enhancement technique, based on a dual descent method, which exhibits better performance in phase preservation compared to independent enhancement methods. Their work involves a coupled optimization formulation that uses a sparsity enforcing penalty term as well as a constraint tying the multichannel images together to preserve the cross-channel information. In addition to independent enhancement, the relative phase between the acquisitions can be degraded due to other factors as well, such as platform location uncertainties, leading to phase errors in the data and defocusing in the formed imagery. The performance of airborne SAR systems can be affected severely by such errors. We ii propose an optimization formulation that combines Ramakrishnan et al.'s coupled IfSAR enhancement method with the sparsity-driven autofocus (SDA) approach of Onhon and Cetin to alleviate the effects of phase errors due to motion errors in the context of IfSAR imaging. Our method solves the joint optimization problem with a Lagrangian optimization method iteratively. In our preliminary experimental analysis, we have obtained results of our method on synthetic SAR images and compared its performance to existing methods. As a second contribution of this thesis, we have developed a software toolbox for end-to-end interferometric SAR processing. This toolbox is capable of performing the fundamental steps of SAR Interferometry Processing. The thesis contains the detailed explanation of the algorithms implemented in the SAR Interferometry Toolbox. Test results are also provided to demonstrate the performance of the Toolbox under different scenarios.

In this paper, we propose a synthetic aperture radar (SAR) moving-target imaging approach that exploits the low-rank and sparse decomposition (LRSD) of subaperture data. The low-rank component consists of the static background whereas the sparse component captures the moving targets. This allows the reconstruction of a full resolution moving target image separate from the static background image after LRSD. Furthermore, it facilitates the applicability of sparsity-driven moving target imaging in low signal to clutter ratio (SCR) scenarios. We demonstrate the effectiveness of our approach with experiments on synthetic as well as real SAR data.

Detecting and capturing 3D human structures from the intensity-based image sequences is an inherently arguable problem, which attracted attention of several researchers especially in real-time activity recognition (Real-AR). These Real-AR systems have been significantly enhanced by using depth intensity sensors that gives maximum information, in spite of the fact that conventional Real-AR systems are using RGB video sensors. This study proposed a depth-based routine-logging Real-AR system to identify the daily human activity routines and to make these surroundings an intelligent living space. Our real-time routine-logging Real-AR system is categorized into two categories. The data collection with the use of a depth camera, feature extraction based on joint information and training/recognition of each activity. In-addition, the recognition mechanism locates, and pinpoints the learned activities and induces routine-logs. The evaluation applied on the depth datasets (self-annotated and MSRAction3D datasets) demonstrated that proposed system can achieve better recognition rates and robust as compare to state-of-the-art methods. Our Real-AR should be feasibly accessible and permanently used in behavior monitoring applications, humanoid-robot systems and e-medical therapy systems.

Image registration is the process of matching, aligning and overlaying two or more images of a scene, which are captured from different viewpoints. It is extensively used in numerous vision based applications. Image registration has five main stages: Feature Detection and Description; Feature Matching; Outlier Rejection; Derivation of Transformation Function; and Image Reconstruction. Timing and accuracy of feature-based Image Registration mainly depend on computational efficiency and robustness of the selected feature-detector-descriptor, respectively. Therefore, choice of feature-detector-descriptor is a critical decision in feature-matching applications. This article presents a comprehensive comparison of SIFT, SURF, KAZE, AKAZE, ORB, and BRISK algorithms. It also elucidates a critical dilemma: Which algorithm is more invariant to scale, rotation and viewpoint changes? To investigate this problem, image matching has been performed with these features to match the scaled versions (5% to 500%), rotated versions (0° to 360°), and perspective-transformed versions of standard images with the original ones. Experiments have been conducted on diverse images taken from benchmark datasets: University of OXFORD, MATLAB, VLFeat, and OpenCV. Nearest-Neighbor-Distance-Ratio has been used as the feature-matching strategy while RANSAC has been applied for rejecting outliers and fitting the transformation models. Results are presented in terms of quantitative comparison, feature-detection-description time, feature-matching time, time of outlier-rejection and model fitting, repeatability, and error in recovered results as compared to the ground-truths. SIFT and BRISK are found to be the most accurate algorithms while ORB and BRISK are most efficient. The article comprises rich information that will be very useful for making important decisions in vision based applications and main aim of this work is to set a benchmark for researchers, regardless of any particular area.

This paper presents spatiotemporal hybrid features, human tracking, and activity recognition into a single framework from video sequences captured by a RGB-D sensor. Initially, we received a sequence of depth maps to extract human silhouettes from the noisy background and track them using temporal human motion information from each frame. Then, hybrid features as optical flow motion features and distance parameters features are extracted from the depth silhouette region and used in an augmented form to work as patiotemporal features. In order to represent each activity in a better way, the augmented features are being clustered and symbolized by self-organization maps. Finally, these features are then processed by hidden Markov models to train and recognize human activities based on transition and emission probabilities values. The experimental results show the superiority of the proposed method over the state-of-the-art methods using two challenging depth images datasets.